Skip to content

Issues: sgl-project/sglang

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

432 Open 1,089 Closed
432 Open 1,089 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

[Bug] Why is the model not stopping and not producing any output? pls check screenshot.
#4659 opened Mar 22, 2025 by seasoncool
4 of 5 tasks
[Bug] Tool Call Parsing for Llama 32 Models Results in Exception
#4658 opened Mar 21, 2025 by kyle-pena-kuzco
5 tasks done
[Bug] test_verl_engine.py fails when CUDA_VISIBLE_DEVICES is not sequential from 0 (invalid device ordinal)
#4657 opened Mar 21, 2025 by zyzshishui
5 tasks done
PD Tracker
#4655 opened Mar 21, 2025 by ByronHsu
1 of 7 tasks
[Bug] Integrate DeepEP into SGLang Multi node case is failed on H100
#4653 opened Mar 21, 2025 by xwz-ol
5 tasks
[Feature] make page size greater than one compatible with EAGLE high priority
#4652 opened Mar 21, 2025 by zhyncs
2 tasks
@merrymercy @ispobock @zhyncs
[Bug] when using --device cpu parameter, running on a Node with GPU, a crash occurs
#4651 opened Mar 21, 2025 by 374660267
2 of 5 tasks
@jhinpan
1
use the same Test Environment and same Engine arguments as #4616 ,but the speed is slow
#4649 opened Mar 21, 2025 by mengli0
1
mpirun using P2P/CUMEM, sglang lauch using P2P/IPC by nccl
#4646 opened Mar 21, 2025 by xiaoxlm
@jhinpan
1
[Bug] Qwen2.5-VL-72B image input not working in SGLang, works fine in vLLM good first issue Good for newcomers
#4645 opened Mar 21, 2025 by qWaitCrypto
5 tasks done
@jhinpan @woodx9
5
[Bug] perf drop when flashinfer update from 0.1.6 to 0.2.x
#4642 opened Mar 21, 2025 by MichoChan
5 tasks done
@jhinpan
1
[Question] Why do not use pinned memory in Hierarchical Radix Cache?
#4636 opened Mar 20, 2025 by lausannel
@jhinpan
1
How to config openai files storage path?
#4635 opened Mar 20, 2025 by debkanchan
@jhinpan
2
[Bug] save_sharded_state failed
#4633 opened Mar 20, 2025 by AllenXu93
3 of 5 tasks
1
@jhinpan
1
[Bug] Docker Image - ModuleNotFoundError: No module named 'aiter'
#4630 opened Mar 20, 2025 by chn-lee-yumi
5 tasks done
@jhinpan
1
[Bug] Gemma-3-27b-instruct use CPU and MEM but not GPU while inference images visIon-LM
#4627 opened Mar 20, 2025 by engchina
5 tasks done
@jhinpan
1
[Bug] When sglang starts, and when inputting around 6000 sequentially, with 10 different input requests, the first response of sglang fluctuates abnormally.
#4626 opened Mar 20, 2025 by whybeyoung
5 tasks done
@shuaills
1
[Bug] DeepEP CUDA Error /sgl-workspace/DeepEP/csrc/kernels/internode.cu:133 'named symbol not found'
#4624 opened Mar 20, 2025 by pc-neo
2 of 5 tasks
@shuaills
1
How to quantify the Qwen2.5-VL-3B model?
#4619 opened Mar 20, 2025 by Gusha-nye
@shuaills
2
[0.4.4.post1] DeepSeek-R1 Optimization Option Ablations high priority
#4616 opened Mar 20, 2025 by m0g1cian
@zhaochenyang20
2
[Feature] MLA remove materialization in weight absorprion
#4615 opened Mar 20, 2025 by edchangy11
2 tasks
@shuaills
3
How to Deploy a Custom Model in SGLang? Is There Any Documentation Available?
#4614 opened Mar 20, 2025 by kanghuazhao
@shuaills
2
[Bug] start Google Gemma3 with SGLang 0.4.4.post1: ValueError: '<class 'sglang.srt.configs.qwen2_5_vl_config.Qwen2_5_VLConfig'>' is already used by a Transformers model.
#4607 opened Mar 20, 2025 by didier-durand
4 of 5 tasks
@shuaills
7
[Bug] sglang decode out of memory
#4602 opened Mar 20, 2025 by ICENacl
5 tasks done
[Bug] When using hlapi, if both stream, beam search and return_log_probs are enabled simultaneously during inference, an error is reported: assert len(output.logprobs) == output.length.
#4601 opened Mar 20, 2025 by shiqingzhangCSU
5 tasks done
ProTip! Add no:assignee to see everything that’s not assigned.